Should we still be optimizing “in the small”?

Question

I was changing my for loop to increment using ++i instead of i++ and got to thinking, is this really necessary anymore? Surely today\'s compilers

Answer 1

Only if you know for certain that they're relevant. This means either you've investigated this issue before on your particular compiler or you've already done the following:

1. produced functional code
2. profiled that code
3. identified bottlenecks
4. simplified design to eliminate bottlenecks
5. chosen algorithms that minimize calls into bottlenecks

If you've done all of these things then often the best thing to do is get your compiler to emit something lower level that you can examine yourself (like assembly) and make specific judgments based on that. In my experience every compiler is a little different. Sometimes an optimization on one causes another to produce larger less efficient code.

If you haven't already done these things then I'd call it premature optimization and I'd recommend against it. Optimizing before doing these things gives rewards which are disproportionately small when compared to the cost involved.